Every uncensored and abliterated model I run, on one key (from $0.084 per million)

Today, 04:11 AM

Uncensored and abliterated builds usually mean a separate host, a separate account and a separate balance for each one. I put them on the same key and base URL as Claude, GPT and Gemini, so switching is a change to the model string.

The models
Per million tokens, input / output:

GLM-5.3 Flash Uncensored glm-5.3-flash-uncensored: $0.084 / $0.084
GLM-5.3 Uncensored glm-5.3-uncensored: $0.18 / $0.18
Roleplay Uncensored roleplay-uncensored: $0.30 / $0.60 (list $1 / $2)
GLM-5.3 Flash Abliterated glm-5.3-flash-abliterated: $0.66 / $0.66
DeepSeek V4 Pro Uncensored deepseek-v4-pro-uncensored: $0.75 / $0.75
GLM-5.3 Abliterated glm-5.3-abliterated: $2.25 / $2.25

Abliterated builds have the refusal direction projected out of the weights. Uncensored builds are fine-tuned instead. They answer differently, so run both on your own prompts before you pick one.

Hooking them up
Anything that takes an OpenAI-compatible endpoint works: SillyTavern (Chat Completion, Custom source), Open WebUI, LibreChat, or your own code.

Code:

from openai import OpenAI



client = OpenAI(base_url="https://api.uplinkswitch.com/v1", api_key="sk-uplink-live-...")

r = client.chat.completions.create(

    model="glm-5.3-uncensored",

    messages=[{"role": "user", "content": "..."}],

)

print(r.choices[0].message.content)

Every request gets a usage record with its token counts and the charge, so a long chat never surprises you. Full list with live prices: uplinkswitch.com/models.

I run UplinkSwitch.